audeering/opensmile

Accept a pcm data, return to extract feature data

srdfjy opened this issue · 2 comments

Hi,We know that when executing SMILExtract, you need to specify an input file and output a file. E.g:
SMILExtract -C config/emobase/emobase2010.conf -I ./1.WAV -O ./1.arff

I want to input a pcm data and return the extracted feature data. Is this possible?

openSMILE comes with SMILEapi which is a C API allowing to embed openSMILE into other applications. Instead of cWaveSource and cArffSink components for file reading/writing, you would then use special components for reading/writing data in memory: cExternalDataSource, cExternalAudioDataSource and cExternalDataSink.

You can find more information on this in the openSMILE documentation in the "SMILEapi" section.